LinkedIn API cursor-based pagination and result set bounding

When retrieving large datasets via the LinkedIn API, the platform implements cursor-based pagination using a start parameter to maintain stability and prevent result duplication during concurrent data updates.

To manage resource consumption, the count parameter is used to bound the number of entities returned per request. However, when dealing with extremely large result sets that may be subject to internal system timeouts or truncation, the behavior of the pagination object becomes critical for ensuring complete data retrieval.

Given these constraints, what is the recommended strategy for handling truncated result sets when the pagination cursor is provided but the total record count is not explicitly returned? How should a client application determine if a result set has been prematurely capped by platform stability limits versus reaching the actual end of the dataset?